Hacker News Reader

Comment by westurner | Hacker News Reader

Notes from "Show HN: Recall: Give Claude memory with Redis-backed persistent context" a few days ago: https://news.ycombinator.com/context?id=45517613 .. https://github.com/joseairosa/recall

How does buildautomata_memory_mcp differ in functionality and implementation from recall, and post-ECAN OpenCog STM/LTM decay?

View on HN

A bit more simple. An MCP server and an interactive command line tool for CLI. It is written in python. Feel free to check it out. The system exposes the tools to the agent to use, it doesn't do any type of context adds automatically. It can link the memory system from any MCP compatible client. So claude desktop and claude code can see the same memories. Or you can use different MCP servers for each instance..just move the folder and update the username/agentname.

I'm working with a loop that keeps reverting and reimplementing the same code and there's not much risk of any malicious input in the context given the chat /save (which doesn't include pytest outputs it parsed for example).

How to detect when this occurs?

"A small number of samples can poison LLMs of any size" (yesterday) https://news.ycombinator.com/item?id=45529587

The version history with diffs helps with that. When you store/update memories, the system tracks changes over time. You can query the timeline to see "what solutions did I already try for X problem?" The Memory decay approach (tracking access patterns + importance decay) also surfaces this - if you keep accessing the same memory about a bug that never gets resolved, that's a signal. Not automatic loop detection yet, but the infrastructure is there. Could build a simple heuristic: if similar memory content gets created/updated N times within short timeframe, flag it as potential loop. The poisoning concern is real. Our design mitigates it somewhat - agents explicitly choose what to store (tool-based, not automatic injection), and you can prune/audit memories.

> Could build a simple heuristic: if similar memory content gets created/updated N times within short timeframe, flag it as potential loop

That would probably prevent waste and could scale.

From https://news.ycombinator.com/context?id=45561039 :

> /? llm firewall: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

There's probably already a name for 'Sensitivity analysis'; how does the output vary with/without particular context?

Reply on news.ycombinator.com